la modellazione dei processi in medicina corso di ......università degli studi di trieste...
Post on 22-Jan-2021
3 Views
Preview:
TRANSCRIPT
Università degli Studi di Trieste
Dipartimento di Ingegneria e Architettura
Corso di Laurea Magistrale in INGEGNERIA CLINICA
LA MODELLAZIONE DEI PROCESSI IN MEDICINA
Corso di Informatica MedicaDocente Sara Renata Francesca MARCEGLIA
PROCESS MODELING
November 10, 2014Sara Marceglia, PhD 2
Model à abstraction used to represent and describe the process under examination
Askari et al, 2013
Ferrante et al, 2013
Lenz & Reichert, 2007
Quaglini et al, 2001
Panzarasa et al, 2002
HEALTHCARE PROCESSES
3
COMPLEXITY
Multiple stakeholders
Multiple systems
Patient’s response to treatment
Patient’s engagement
Patient’s empowerment
Evolving knowledge
Evidence based practices
Multiple action domains
COMPLEXITY AND TRANSPARENCY: CLINICAL ISSUES
November 10, 2014Sara Marceglia, PhD 4
DOMAIN FEATURE ISSUE
Med
ical
kno
wle
dge Evidence based medicine, guidelines,
recommendationsDynamic Evolution of medical evidence-
based knowledgeLocal practices Flexibility to include the local context
Clinician’s personal experience, habits and skills
Multiple stakeholders interaction
Learning by practice Learning curves that interfere with timing
Building evidence from practice Process mining with big data analysis
Pers
onal
izat
ion
and
Res
pons
e to
trea
tmen
t Time - immediate response or long-term response
Uncertainty
Compliance - depending on patient’s engagement
Indeterminacy
Expected outcomes Definition of outcome variables
Patient’s feedback to treatmentUncertainty due to patient’s reported
outcome measures
Patient-centric approach Exceptions and Flexibility
COMPLEXITY AND TRANSPARENCY: ORGANIZATIONAL ISSUES
DOMAIN FEATURE ISSUE
Syst
ems a
nd
infr
astr
uctu
res Data exchange among different systems Technological Interoperability
Introduction of local IT infrastructural constraints
Flexibility to include localcontext
Compliance with standards Technological Interoperability
Act
ors
Related actions, multiple profiles executing the same task
Multiple responsibilities Transparency
Data access Data Protection
Domain knowledge of the specific profile Semantic Interoperability
EXPECTATIONS FROM HEALTHCARE PROCESS MODELING
6
CLINICAL EXPECTATIONS
• Establishing shared protocols for patient’s care.
• Facilitating adherence to the shared protocols, thus limiting problems due to incomplete communication or misunderstandings among different actors, ultimately increasing patient’s safety.
• Monitoring deviances from the protocols, redundancies, and failures, thus early identifying problems that could lead to un-prevented errors.
ORGANIZATIONAL AND TECHNOLOGICAL EXPECTATIONS
• Fully understanding the information flow, thus identifying requirements and specifications for information system re-engineering and interoperability.
• Detecting process weaknesses thus designing corrective measures.
• Optimizing the use of resources.
RESULTS OF PROCESS MODELING: e-PRESCRIBING
Marceglia et al, Methods Inform Med, 2013
The Assign Phase
The Administer Phase
RESULTS OF PROCESS MODELING: e-PRESCRIBING (2)
UNDERSTANDING WHAT IS NEEDED
•None of the e-prescribing systems studied manages the drugadministration phase à new tools for the safe and monitored drug administration at home
•Only one e-prescribing systemprovides some support duringprescription à new systems to integrate drug references and drug-drug interactions for GPs
EVALUATING BENEFITS AND COMPARING EXISTING SYSTEMS
•If a phase of the e-prescribing process is completed according to the model, there are benefits for the healthcare system à Quality, Efficiency, Access
•Implemented systems can be represented through the modelled functionalities à comparing the functionalities can be translated into comparing the benefits
Marceglia et al, Methods Inform Med, 2013
RESULTS OF PROCESS MODELING: ONCOLOGY
The ambulatory unit provides pharmacological anticancer therapy and supportive therapy to oncological patients
RESULTS OF PROCESS MODELING: ONCOLOGY
LOCAL LEVEL
• Definition of the specifications for a new module of the hospital information system able to manage the information loss during the ambulatory process
TRANSLATIONAL LEVEL
• Definition and representation of the care pathway of the oncologic patient during ambulatory anticancer therapy can be used in other oncologic settings
RESULTS OF PROCESS MODELING: INTEGRATED HOMECARE
Homecare services for residents (care, palliative
care, rehabilitation)
RESULTS OF PROCESS MODELING: INTEGRATED HOMECARE
Sara Marceglia, PhD 12
TRANSPARENCY •The whole process of integrated homecare in the specific environmental setting was fully evaluated
•All the health information systems for patient’s data collection and management were analyzed in terms of functionalities and information processed
PROCESS EVALUATION •The most critical issues in the process were identified•The redundancies and duplications of information across different systems were mapped
PROCESS OPTIMIZATION•Possible solutions to the critical issues were identified and introduced in the model•The model can be used to define measures of the efficiency of the service represented
Mancanza terminologia standard per attività, patologie e dispositiviMancata corrispondenza con PO per frequenza di interventiImpossibilità di legare dispositivi e PO
Inserimento ICD-9 per patologieDistinzione tra dispositivi temporanei e permanenti da inserire in POUtilizzo codici ministeriali per i dispositivi
MODELING METHODOLOGY OVERVIEW
DOMAIN ANALYSIS
FORMAL NOTATION
SELECTION
DEFINE CRITICAL
ISSUES PRE-
MODELING MODELING METRICS DEFINITION
MAPPING TO
SPECIFIC LANGUAGE
ENVIRONMENTAL CONTEXT ANALYSIS
CONCEPTUAL MODELING
LOGICAL MODELING
PROCESS ANALYSTS
DOMAIN EXPERTS
Modeling languages
Requirement analysis
Structured interviews Guidelines
Evidence-based practice
On-the-field experience
EXPERT VALIDATION
Domain knowledge
DOMAIN ANALYSIS
• Identifying and analyzing the available sources of information to fully understand the domain of interest
• Evidence-based knowledge à international guidelines and recommendations
• Local domain à– local practices– specific clinical pathways already in use locally – focus groups and interviews to the medical staff
and/or the patient/caregivers, to highlight the personal experience of the actors involved in the process
• Description of the information systems already in use àhelps planning the model deployment in the real everyday practice.
DOMAIN EXPERTS
SELECTION OF THE FORMAL NOTATION: UML
ANALYST
“The Unified Modeling Language (UML) is a graphical language for
specifying, visualizing, constructing, and documenting
the artifacts of software systems, as well as for business modeling
and other non-software systems”.
STATIC VIEWS DYNAMIC VIEWS
Class Diagram
Use-case Diagram
Component Diagram
Activity Diagram
Sequence Diagram
UML: ADVANTAGES
• Graphical language à
• Enables the communication between domain experts and analysts
• Easy to understand by non-experts of computer sciences
• Provides different views on the processà
• Static and dynamic diagrams
• Structural, Behavioural, and Interaction diagrams
• The final UML model can be used as specification for the development of an IT system
• The use of UML promotes modularity and facilitates future changes
• It allows the evaluation of the whole system, also at the deployment level
PRE-MODELING
• Provide a high-level process description (process phases)– Functional aspects (main activities of the process, objects and
data items managed)– Organizational aspects (agents, roles, skills, availabilities,
authorizations required to enact the process) � understand actors’ responsibilities on the main activities business aspects (goals to be achieved).
• Define a list of goals of the process modeling effort
• Identification of outcomes and integration with patient-reported outcome measures
MODELING
• The modeling step starts from the previously collected information and produces a conceptual model of the process according to the formal notation adopted
• The conceptual model comprises:– the schema of the process– its variables– the specification of the expected exceptions– the specification of the transactions– the access control model– the description of the interactions with the external information
system. • The results of the modeling are then validated by the
domain experts, before continuing with the design
MODEL EVALUATION
Assess if the model is “syntactically correct”•Verify the internal consistency of the model.•Assess its usability as a starting point for the logical modeling
towards the software implementation.
Assess if the model is “semantically correct”• Iterative approach based on focus groups or external
validators.•Analysis by actor type to validate the flow of information in
the simplest activities of the process.•Analysis with multiple actors to test the whole model. • The model is final when a total agreement between experts
and analysts is reached.Askari et al. 2013
ANALYST DOMAIN EXPERTS
METRICS AND GQM
• METRIC = a quantitative measure of the degree to which a system, component, or process possesses a given attribute
• Metrics for the evaluation of health ITs cannot be directly derived from the model itself à the model can be the basis for identifying the outcome variables to be introduced into e-management techniques as metrics for evaluation.
• Goal Question Metrics (GQM) à
– allows selecting metrics with a top-down and goal-oriented approach
– the identification of the metrics starts from the definition of goals
– The definition of the goals is done during the conceptual modeling phase
• GQM has three levels à
– Goal: Conceptual level, defines the main purposes of a work to be measured;
– Question: Operational level, defines a set of questions useful for achieving the goals;
– Metric: Quantitative level, defines a set of metrics for answering the questions in a measurable way.
MODEL IMPLEMENTATION
• Mapping to a specific executable language
• Building a system implementing the model
• Create new modules of an already existing system to implement the process
top related